Simons Foundation Lectures

Simons Foundation Lectures

By Simons FoundationScienceEducationMathematics
Download on the App Store

All Simons Foundation Lectures episodes

  • Matthieu Dartiailh - Phase signature of topological transition in Josephson juncts (Oct 11, 2019)

    Topological transition transforms common superconductivity into an exotic phase of matter, which holds promise for fault-tolerant quantum computing. A hallmark of this transition is the emergence of Majorana states. While two-dimensional semiconductor/superconductor heterostructures are desirable platforms for topological superconductivity, direct phase measurements as the fingerprint of the underlying topological transition are conspicuously missing. On gate tunable Josephson junctions made on epitaxial Al/InAs, we observe a closing and a reopening of the superconducting gap with increasing in-plane magnetic field. Since our junctions are embedded into a phase-sensitive SQUID, we are able to measure a π-jump in the superconducting phase across the junction coincident with the closing and reopening of the superconducting gap. Theoretical simulations confirm this transition is topological and compatible with the emergence of Majorana states while the magnetic field angle dependence of the transition further constrain this scenario. Remarkably, in each junction, this topological transition can be controlled by changing the gate voltage. These findings reveal versatile two-dimensional platforms for scalable topological quantum computing.
    More details: https://www.simonsfoundation.org/event/qclub-seminar-matthieu-dartiailh/

    Uploaded 22-11-19

    Learn more about your ad choices. Visit megaphone.fm/adchoices

    41 min
  • Matthieu Wyart: How Much Data Is Required to Learn the Structure of Language? (September 25, 2026)

    Despite the fact that experimental neural scaling laws have substantially guided empirical progress in large-scale machine learning, no existing theory can quantitatively predict the exponents of these important laws for any modern LLM trained on natural languages. It is also unclear whether they can be beaten by training with other objectives. Matthieu Wyart will review the efforts of the collaboration in studying this phenomenon, including the first such theory in the case of data-limited scaling laws. Wyart will then introduce synthetic languages displaying a hidden hierarchy (a central notion in both linguistics and computer vision), and discuss how these languages are learnt by deep, but not shallow, nets. Scaling laws in these synthetic languages follow from the same principle as in natural languages when trained at the token level, but these languages can be learnt much more efficiently by machines predicting in an abstract space — justifying the success of joint embedding predictive architectures (JEPA) and other machines learning from their own latents.

    Uploaded 06-11-26

    For more information please visit: https://www.simonsfoundation.org/events/simons-collaboration-on-the-physics-of-learning-and-neural-computation-annual-meeting-2026/

    Learn more about your ad choices. Visit megaphone.fm/adchoices

    59 min
  • ICM 2026 Invited Section Lecture - Urmila Mahadev

    Classical Control of a Quantum System

    In this talk, we'll discuss the connection between the problem of learning with errors (LWE) and quantum computation. LWE is a widely studied cryptographic assumption which is believed to be secure even for quantum computers, unlike assumptions that are commonly used today. We'll discuss how, even though LWE does not seem to have enough structure to be amenable to quantum algorithms, it does have enough structure to enable uniquely quantum functionality. We'll illustrate this by describing how the LWE assumption can be used to classically control a quantum system.

    Uploaded 05-10-26

    Learn more about your ad choices. Visit megaphone.fm/adchoices

    52 min
  • ICM 2026 Invited Section Lecture - Emily Riehl

    Synthetic Perspectives on Higher Structures

    Through painstaking work, higher structures, such as weak infinite-dimensional categories, can be built "analytically" out of sets. Elaborate formalisms have been developed to ensure that constructions involving these objects are equivalence-invariant. We present an alternate approach using domain-specific formal languages to work with higher structures "synthetically." We argue such formal systems make definitions, theorems, and proofs both easier to understand and easier to formalize. Our contributions to the synthetic theory of higher categories were developed over the course of joint work with Dominic Verity, Mike Shulman, Evan Cavallo, and Christian Sattler, among others.

    Uploaded 07-10-26

    Learn more about your ad choices. Visit megaphone.fm/adchoices

    50 min
  • ICM 2026 Invited Section Lecture – David Witt Nyström

    A Calabi-Yau Theorem for Degenerations of Compact Kähler Manifolds

    To understand a compact Kähler manifold X, it is often helpful to study its degenerations. The idea of non-Archimedean (n-A) Kähler geometry is to think of such degenerations as (possibly weak) Kähler potentials on an associated n-A space. The n-A Monge-Ampère measure of a degeneration is then supposed to encode how the volume of X distributes as X degenerates. In this talk I will discuss a Calabi-Yau theorem in this setting, first proved in the algebraic case by Sébastien Boucksom and Mattias Jonsson (building on earlier work of Sébastien Boucksom, Charles Favre and Mattias Jonsson), and subsequently proved in the general Kähler case by Pietro Mesquita-Piccione and myself. The proof in the Kähler case relies on properties of big cohomology classes, and is related to the transcendental Morse inequality conjecture of Sébastien Boucksom, Jean-Pierre Demailly, Mihai Păun and Thomas Peternell.

    Uploaded 06-10-26

    Learn more about your ad choices. Visit megaphone.fm/adchoices

    51 min